[time-nuts] Chinese Scopes

Jim Lux jimlux at earthlink.net
Tue Apr 17 14:34:28 UTC 2012


On 4/17/12 7:15 AM, Robert Darlington wrote:
> I need lots of memory on scopes.  A buddy of mine I worked with in the
> ultrasound world actually yelled at the Tek product management and
> asked if they actually *use* oscilloscopes.  The answer was a sheepish
> no, and yet they felt qualified to develop the products for the
> company.
>
> The cheap Aktakom scope I have has plenty.  10 million samples (you
> can select less if you want) and will write out to usb thumb drives.
> It's definitely a toy scope with lots of noise, but it's useful for
> some things.
>
> What we do is send out pulses or chirps and look at what returns.
> There are tens of millisecond delays between what we send out and what
> we receive and the echos.    With traditional low memory scopes we
> simply can't get by.  Thankfully Tek is learning that memory is cheap
> and 2500 samples was hardly sufficient in the 70s, let alone now!


Yes, just like in the radar world (really, ultrasound and radar are 
really similar.. same kinds of pulse compression and signal processing)


Back in 1998-1999, I was buying digitizer cards from Gage Applied 
Sciences (since acquired by Tek, as it happens), and one of their big 
markets was for ultrasound.  Same for Signatec (another mfr of fast 
digitizer cards for PCs)

Another case where deep memory is nice is when you don't know exactly 
when the signal is going to arrive, it's very low SNR, so you want to 
record a long time, and then go look for the signal later.  But that's 
more a data capture problem than a bench oscilloscope problem.

say you were recording off-the-air GPS signals.  You want to record a 
couple milliseconds, at least (so you get at least 1 code epoch), and 
you need to record at least 10 MHz bandwidth.  That's only, say, 64,000 
samples, but you might want to record a whole 50 bps nav message bit, so 
then you need ot record 40-50 milliseconds, and the record length starts 
to grow.

Again, that's more of a data recording problem than an oscilloscope problem.

It's the wideband pulsed waveforms where you want to compare pulse to 
pulse is where deep memory in an oscilloscope is nice.  The digitizer 
cards are ok, but "real oscilloscopes" tend to have better input 
amplifiers and such.






More information about the time-nuts mailing list